Workflow resume and retry

At a glance

When a workflow fails, you do not need to restart everything from the beginning. Checkpoint resume lets you pick up execution from the last successful block. Configure automatic retry policies to handle transient errors (network timeout, temporarily unavailable service) without manual intervention.

Before you begin

• A workflow must be created and have been executed at least once.
• The execution must be in durable mode (Run button, not the Test button).

Steps

Resume from a checkpoint

When a workflow fails during execution:

1. Open the workflow History.
2. Identify the failed execution (red status).
3. Click Resume.
4. The workflow restarts from the last successful block, without re-executing previous blocks.

Example: a 10-block workflow fails at block 7 (temporarily unavailable external API call). Resuming re-executes only blocks 7 through 10, preserving the results of blocks 1 through 6.

tip

Resuming is faster and safer than a full re-execution, because the side effects of already-executed blocks are not repeated (emails not resent, data not duplicated).

Configure automatic retry

1. Open your workflow in the Workflow Builder.
2. Click the Settings icon in the toolbar.
3. In the Resilience section, configure:

Parameter	Description	Default value
Retry enabled	Enable automatic retry on failure.	No
Max attempts	How many times to retry before giving up.	3
Delay between attempts	Increasing wait time between attempts.	Exponential (1s, 2s, 4s...)

4. Save.

View attempts

1. In the history, each execution displays the number of attempts.
2. Click an execution to see the details:
   • Attempt 1: failed (network error)
   • Attempt 2: failed (timeout)
   • Attempt 3: succeeded

Manage the error queue

Executions that fail after all attempts are placed in the error queue (dead letter):

1. Open the Error queue section in the workflow history.
2. For each entry, you can:
   • Retry: manually relaunch the execution.
   • Discard: remove the entry from the queue.
3. Review the error message to diagnose the cause.

Cancel a running workflow

To stop a workflow during execution:

1. In the history, identify the running execution (blue status).
2. Click Cancel.
3. Confirm the cancellation.
4. The workflow transitions to cancelled status. The current block is interrupted.

Expected outcome

Your workflows are resilient to transient errors. Failures are automatically retried according to the configured policy, and failed executions can be resumed from the last checkpoint without re-executing previous blocks.

Limitations and common errors

Situation	Solution
Resume re-executes everything from the start	Verify that the execution was launched in durable mode (Run button) and not in test mode.
Automatic retry does not trigger	Verify that retry is enabled in the workflow settings.
Execution still fails after 3 attempts	The problem is probably not transient. Check the failing block and fix the root cause.
The error queue is full	Sort the entries: retry recoverable cases and discard obsolete ones.

Need help?

Contact us: Support and contact.